[TenTec] Noise Reduction Setting

[Dr. Gerald N. Johnson]

On Fri, 2006-12-08 at 22:28 -0500, Gary Hoffman wrote:
> Indeed yes, all correct.  This is one of the "other" methods I had in mind
> when I wrote one of my earlier posts.  And narrow bandwidth makes the job of
> separating "good" signals from "bad" signals more difficult when using this
> particular method.  Of course there are ways to mitigate that also.  Takes
> more processing though.
> 
> Another point...the signal is highly correlated.  Hence little if any
> reduction takes place really close to the signal.  The noise further from
> the signal is, we hope, highly uncorrelated.  Hence it gets reduced a ton.
> Result ?  The signal level away from the "good" signal drops precipitously,
> while that close in does not - giving the appearance of a band pass filter
> having been dropped in place over the signal.  Can be kinda misleading huh ?
> 
> 73 de Gary, AA2IZ
> 
Yes the working autocorrelator has an output spectrum a lot like a
narrow bandpass filter with flat skirts not all that far down. The noise
reduction depends on the noise being truly random (which is not true for
lighting static, computer hash, and ignition noise, but only for circuit
noise which is not a factor at frequencies below a GHz these days, maybe
12 GHz with 20 K RF stage noise factors) and the bandwidth of the input
noise. The difference works better when the noise bandwidth is broad.
Feeding white random noise through a narrow filter gives it more
correlation that it had when not filtered. Doesn't matter if its an
analogue or a DSP filter. It does matter if the filter rings like S-line
filters then the characteristic frequency of the filter shows up in the
output spectrum.

Allow me to propose a test of the difference between autocorrelation and
a narrow filter as the noise reduction process. Check a frequency with
multiple weak signals, maybe a multiple tone HF data link. Note that if
the NR works it enhances all the signals to noise while a narrow
tracking filter would enhance only ONE of those signals. Tracking a
dozen tones with filters would take a lot more DSP processing, and I
think far more than it has.

And another test, I believe that a continuous carrier will get better
S/N than a keyed signal. It doesn't have the gaps in the signal.


-- 
73, Jerry, K0CQ,
All content copyright Dr. Gerald N. Johnson, electrical engineer